Complex Applications of Hierarchical Linear Modeling in Studies of Science and Mathematics Achievement

• In Event: 46.086 - Roundtable Session 16
  In Roundtable Session: 46.086-10 - Studying Achievement in Math and Science

Abstract

Analyses of factors impacting science and mathematics achievement using hierarchical linear modeling (HLM) are common in education. Traditional applications of HLM to nested cross-sectional data (e.g., students nested within schools), as well as longitudinal data, are well known. However, these models are inadequate in several research settings that require more complex models to ensure the validity of inferences about relationships of interest. In this paper we describe two unrelated categories of models and illustrate each with empirical data and the steps needed to use widely available data analysis software to perform the analyses. Our goal is to increase the understanding of, and accessibility to, these models for a broad audience of researchers studying factors related to science and mathematics achievement.

Authors

• Michael R. Harwell, University of Minnesota

• Mario Moreno, University of Minnesota

• Selcen Guzey, Purdue University

• Alison Phillips, University of Minnesota - Twin Cities

• Tamara Jo Moore, Purdue University